Electronic copy available at: http://ssrn.com/abstract=1316214

The Value of Shareholder Activism: New

Evidence from the Split-share Structure

Reform in China

Li Liao� Meijuan Shi Hao Wang

November 24, 2008

Abstract

We examine shareholder activism in the recent split-share structure reform in China.

This unique event allows us to avoid the de�ciencies in determining activism proxies and

in measuring their e�ectiveness that plague the previous literature. We �nd that the

e�ectiveness of shareholder activism is in uenced by segmented institutional ownerships.

Particularly, shareholder activism could have both positive and negative impacts on

managerial behaviors in the presence of institutional investors' interest con icts. We

show that, even in its primitive stage, Chinese shareholder activism has demonstrated

positive value in corporate governance.

JEL Classi�cation: G12, G18

Keywords: Shareholder Activism, Split-share Structure Reform, Consider-

ation, Positive Promises

�Li Liao, Meijuan Shi and Hao Wang are all from the School of Economics and Management, Tsinghua Uni-

versity and the Chinese Center of Financial Research (CCFR), Beijing 100084, China. Contact Information:

Tel: 86 10 62797482, Fax: 86 10 62784554; E-mail: liaol@sem.tsinghua.edu.cn.

1

Electronic copy available at: http://ssrn.com/abstract=1316214

1 Introduction

Shareholder activism constitutes an important form of shareholders' e�orts to in uence cor-

porate governance. In recent years, Chinese institutional investors have been increasingly

involved in corporate governance and �rm management as they grow to hold more securi-

ties.1 They are considered a monitoring and disciplinary mechanism of growing importance

in the Chinese �nancial system, in which takeovers are systematically discouraged and boards

are ine�ective (Allen et al: (2005)). In this paper, we take advantage of a special event { the

Chinese split-share structure reform { to examine the value of shareholder activism. This

unique event, in nature, allows us to avoid the de�ciencies in determining activism proxies

and in measuring their e�ectiveness that plague the previous literature.

The split-share structure has existed since the inception of the Chinese A-share market

in the early 1990s.2 The original purposes of the split-share structure were to enable state-

owned enterprises to raise capital and, simultaneously, to allow the Chinese government to

retain control of those enterprises. However, it was widely criticized for fostering specula-

tions and agency problems and for hindering M&A activities (Hwang et al: (2006) and Liao

et al: (2008)). In this split-share structure, two classes of domestic A shares with otherwise

identical features, tradable and non-tradable, coexist for one company. Transactions of the

non-tradable shares are contract-based and subject to approval of the State-owned Assets

1According to the China Securities Regulatory Commission Annul Report 2007, institutional investors

held approximately 50 percent of tradable stocks by the end of year 2007.2A Chinese company that is publicly listed on domestic exchanges may issue three types of shares. A-shares

are domestic common stocks traded on Shanghai Stock Exchange or Shenzhen Stock exchange. B-shares are

listed on domestic exchanges but priced in US dollars. H-shares are listed on Hong Kong Stock Exchange.

2

Supervision and Administration Commission. Approximately 2/3 of the A-shares outstand-

ing were non-tradables mainly held by the Chinese government and its a�liates, who assume

signi�cant control of the �rms. Tradable shares were largely owned by institutional and

individual shareholders. The split-share structure reform was initiated by the Chinese gov-

ernment to convert non-tradable A-shares to tradable shares in an e�ort to revitalize the

Chinese stock market, which has been bearish since its initial partial share issue privatization

(SIP) in 1990s (Liao et al: (2008)).

Since the original costs of non-tradable shares were signi�cantly lower than the market

prices of tradable shares, non-tradable shareholders, represented by �rm management, were

required by the China Securities Regulatory Commission (the Chinese counterpart of the SEC

in the United States, hereafter CSRC) to negotiate with tradable shareholders to deliver a

compensation package to the latter. A reform plan will only be passed if it contains a

compensation package that is approved by at least 2/3 of tradable shareholders. Institutional

shareholders, who held more than 20 percent of the total tradable shares in 2006 (Source:

the WIND database), will play an important role in compensation package negotiations and

approval.

In the same spirit as Woidtke (2002), we use the percentage of tradable shares held by

institutional shareholders as a proxy for shareholder activism. The larger stake institutional

shareholders have in a �rm, the more incentive they will have in monitoring and getting

involved in �rm management. We propose three sets of measures for the e�ectiveness of

shareholder activism on in uencing managerial behaviors and �rm values: consideration,

positive promises made to tradable shareholders and abnormal stock returns around reform

plan approval date. Consideration is measured as the number of shares that non-tradable

shareholders transferred to tradable shareholders in exchange for the trading right for their

3

non-tradable shares. Consideration and positive promises will capture the in uence of insti-

tutional shareholder activism on managerial decisions, while abnormal stock returns around

reform plan approval date will re ect the impacts of activism on �rm performance and �rm

value.

Institutional shareholders face potential con icts of interests in monitoring �rm manage-

ment. Therefore, shareholder activism may have opposite e�ects on managerial decisions.

Pound (1988) propose three hypotheses that predict contradictory relationships between �rm

performance and institutional ownership. We �nd that the impacts of shareholder activism

on �rm management are in uenced by segmented institutional ownerships of tradable shares.

Particularly, The value of consideration tends to be higher when institutional shareholders

hold a larger percentage of tradable shares. In this case, shareholder activism plays a positive

and e�ective role in a�ecting managerial decisions for tradable shareholders' interests. How-

ever, considerations turn to be relatively lower when institutional shareholders own a small

percentage of the �rms' tradable shares. Evidence suggests that institutional shareholders get

actively involved in �rm decision making to pursue di�erent agendas in exchange for private

bene�ts, rather than seeking higher compensation values. We demonstrate that shareholder

activism could have both positive and negative impacts in the presence of institutional share-

holders' interest con icts.

Our analysis in general supports the argument that shareholder activism helps to improve

�rm performance and value. Positive promises made by �rm management, though containing

no direct monetary values, could lead to potential increases in �rm and equity values. We

�nd that the number of positive promises is positively correlated to institutional sharehold-

ing. In addition, the result reports both economically and statistically signi�cant positive

relationships between abnormal stock returns and institutional shareholding.

4

Institutional shareholders may take three types of actions when they are unsatis�ed with a

�rm's management. They could choose to vote with their feet - selling their stocks of the �rm.

They may get actively involved in corporate governance to in uence managerial behaviors,

as suggested by activism. Alternatively, they may choose to remain silent. Hirschman (1971)

summarize those shareholder actions into three categories: exit, voice and loyalty. The

previous literature on examining shareholder activism use shareholder proposals as proxies

for shareholder activism. Mixed results were found on shareholder activism's impacts on

corporate governance and �rm performance. Among others, Smith (1996), Strickland et

al: (1996), Carleton et al: (1998) and Del Guercio and Hawkins (1999) �nd evidence that

supports the success of shareholder activism, while Karpo� et al: (1996), Wahal (1996) and

Gillan and Starks (2000) �nd the opposite.

The disagreement among the previous studies arises from the complexities in selecting

shareholder activism proxies and from the di�culties in measuring their e�ectiveness (Kar-

po� (2001) and Gillan and Starks (1998)). For instance, it is hard, if possible, to assure that

the changes in �rm performance are entirely driven by shareholder proposals. Other events,

such as changing board members and modifying corporate charters, may a�ect �rm perfor-

mance and value as well. Furthermore, changing corporate management may not necessarily

lead to changes in �rm performance. An additional complication is that institutional share-

holders usually negotiate with �rm management privately, rather than publicly, to in uence

�rm decisions. To solve the problems, researchers consider institutional shareholder owner-

ship as a proxy for shareholder activism. Institutional investors would have more incentives

in monitoring and in uencing management, in both public and private ways, if they have

greater interest in a �rm. Carvell and Strebel (1987) and Edelman and Baker (1990) �nd the

"neglected �rm e�ect" { �rm values are negatively correlated with institutional ownership.

5

McConnel and Servaes (1990) show a negative relationship between Tobin's Q and institu-

tional ownership. However, their static analysis remains silent on the key question whether

institutional investors take actions to in uence �rm performance.

The compensation package negotiations in the split-share structure reform o�er an un-

paralleled opportunity to examine the value of shareholder activism. In this study, the dates

on which tradable shareholders and non-tradable shareholders reach agreements on com-

pensation packages are known, whereas the previous literature su�ers di�culty in accurately

pinning down shareholder proposal time, which is not publicly available. During the proposal

time, other corporate events may occur and contaminate the measurement of the e�ective-

ness of shareholder activism. In addition, shareholder proposals contain no legal restrictive

power to �rm management. Institutional investors usually talk with �rm management before

submitting their proposals. If an agreement has been reached, institutional shareholders will

not submit a proposal at all. Thus, shareholder proposal is usually regarded as a result of

failed talk between institutional shareholders and management. Our investigation does not

su�er such problem since management could not ignore institutional shareholders' proposal

because they need institutional shareholders' support for reform plan approval. The split-

share structure reform was market-wide, involving almost all publicly listed companies in

the Chinese market regardless their characteristics. That naturally eliminates the problems

of representativeness and size bias. Moreover, there is no self-selection bias because the re-

forms were mandatory. This event study, in nature, avoids the endogeneity problem that

plagues the previous studies on the relationship between �rm performance and institutional

ownership.

The remainder of this paper is structured as follows. Section 2 introduces research method-

ology and data; Regression analysis is carried out in Section 3; Section 4 performs robustness

6

check; Section 5 concludes the paper.

2 Analytical Framework

2.1 Hypothesis Development

Pound (1988) propose three hypotheses that o�er theoretical predictions on the relationships

between compensation value and institutional ownership. The relationships in turn provide

implications for the in uence of institutional shareholder activism on managerial decisions

and �rm performance.

The e�cient monitoring hypothesis states that the informational and technical advan-

tages possessed by institutional shareholders enable them to better negotiate for higher com-

pensation values than small shareholders. It suggests a positive relationship between institu-

tional ownership and compensation value. The con ict of interest hypothesis predicts that

institutional shareholders may choose to support �rm management decisions in exchange for

favorable relationships and private bene�ts. According to the strategic alignment hypothesis,

institutional shareholders have incentives to cooperate with �rm management for long term

bene�ts. The cooperation in general weakens the monitoring function of institutional share-

holders on �rm management. Both the con ict of interest and strategic alignment hypotheses

predict a negative relationship between institutional ownership and compensation values to

tradable shareholders.

Importantly, those hypotheses predict that there may exist a nonlinear relationship be-

tween compensation package value and institutional ownership. Institutional shareholders

could be more likely to engage in cooperation with �rm management in exchange for private

7

bene�ts that out-weight their potential gain in higher compensation value when their own-

erships are low. In this case, the compensation value is negatively related to institutional

ownership. On the other hand, when institutional shareholders hold a large percentage of

tradable shares, they are more likely to pursue for higher compensation value in reform ne-

gotiations. As a result, compensation value should be positively correlated with institutional

shareholder activism. We propose the following hypothesis to be tested:

Hypothesis 1: There exists a U-shape nonlinear relationship between compensation value

and shareholder activism.

Firm management could make positive promises to sweeten a compensation package in

reform negotiations. Those promises in general are bene�cial to the improvements of �rm

performance and value in the future. For instance, controlling non-tradable shareholders'

promise to inject additional capital signals that there are potential investment opportunities

of positive net present values. Large non-tradable shareholders' promise to increase their

holdings when stock price falls reduces the interest con icts between original non-tradable

and tradable shareholders. Firm management's promise for future dividend payments helps

to mitigate the management-shareholder agency problems by paying out extra cash ows. Xu

et al: (2008) �nd that abnormal stock returns are positively correlated with positive promises

made by �rm management in the split-share structure reform. Given institutional sharehold-

ers' superior capability in compensation negotiations compared to individual shareholders,

we propose and test the following hypothesis:

Hypothesis 2: There exists a positive relationship between the probability of �rm management/non-

tradable shareholders making positive promises and shareholder activism.

Feng and Xu (2006) report an average of 3% abnormal stock returns on reform plan

approval days. The magnitude of abnormal stock returns will re ect the aggregate market

8

expectations on how shareholder activism a�ects �rm future values. Following the e�cient

monitoring hypothesis, we expect a positive relationship between the abnormal returns and

shareholder activism. On the other hand, the con ict of interest and strategic alignment

hypotheses suggest a negative relationship. Thus, we propose and test a third hypothesis:

Hypothesis 3: There exist positive relationships between cumulative abnormal returns

around reform plan approval and shareholder activism.

2.2 Data

Our data on compensation packages, including considerations and promises, of the split-

share reform comes from the WIND database, which is widely used by major research and

�nancial institutions. Firm characteristics, stock returns and institutional shareholding data

are obtained from the Tsinghua University Data Center and the WIND database respectively.

Our original data contain 1223 �rms that completed their split-share structure reforms as of

30 June 2007. We eliminated 11 �nancial �rms, who themselves are institutional shareholders.

We further excluded 75 �rms that were publicly listed less than two years before their reforms

to improve the reliability of our measures of �rm characteristics. Moreover, 42 �rms with

incomplete �nancial information were excluded. Our �nal sample contains 1095 �rms, among

which 165, 859 and 71 �rms �nished their reforms in year 2005, 2006 and 2007 respectively.

There are 30 �rms selected by the CSRC as pilot �rms to carry out split-share reforms in

year 2005.

In our sample, 65% �rms are listed on Shanghai Stock Exchange. The rest of 35% �rms

are listed on Shenzhen Stock Exchange. The average book values of the �rms are RMB 3.66

billion with a minimum of RMB 90 million and a maximum of RMB 521 billion (RMB 7 equals

9

USD 1 approximately). The average leverage ratio is approximately 50%. The average return

on equity is approximately 4%. In our sample, there are approximately 9% �rms that issue

B-/H-shares besides domestic A-shares. Approximately 72% of the �rms are state-controlled,

providing a lower boundary for the number of �rms for which the negotiations between non-

tradable shareholders and tradable shareholders are actually the negotiations between �rm

management and tradable shareholders. The diversi�cation of the �rm characteristics re ects

the fact that our sample includes almost all publicly listed companies in the Chinese stock

market, con�rming that our analysis is immune to self-selection bias.

2.3 Methodology

We follow Woidtke (2002) to use the percentage of tradable shares held by institutional in-

vestors as a proxy for shareholder activism. The idea behind our choice of activism proxy is

that the larger stake institutional shareholders have in a �rm, the more incentive they will

have in monitoring and getting involved in �rm management. We use three sets of explana-

tory variables to measure shareholder activism's impacts on managerial decisions and �rm

performance in our regression analysis. They are consideration (CONS), positive promises

made by �rm management (PPs) and cumulative abnormal stock returns around reform

plan approval date (CARs). We will formally introduce those variables below. Particularly,

consideration and positive promises capture the impacts of activism on managerial decisions.

Abnormal stock returns re ect the aggregate market expectations on the e�ects of activism

on the improvement of �rm future performance and value. Our prior is that positive rela-

tionships between the explanatory variables and institutional ownership would suggest that

the presence of shareholder activism has positive in uence on managerial behaviors and �rm

10

performance, ceteris paribus.

We use the following regressions to analyze the impacts of shareholder activism on con-

sideration and cumulative abnormal returns:

CONSi = �+ �1 � PINi + �2 � PINSQi +nXj=1

�i;j � Controli;j + "i (1)

CARi = �+ �1 � PINi +nXj=1

�i;j � Controli;j + "i (2)

where subscript i denotes �rm i. PIN denotes the percentage of tradable shares held by

institutional shareholders. PINSQ denotes the square of the percentage of tradable shares

held by institutional shareholders. We include PINSQ to capture the potential non-linear

relationship between activism and its impacts on consideration. Controli;j denotes the jth

control variable for �rm i.

Consideration represents the number of non-tradable shares transferred to tradable share-

holders for every ten tradable shares they hold. In order to gain trading right for their non-

tradable shares, non-tradable shareholders need to transfer some of their non-tradable shares

to tradable shareholders of the same �rm as compensation, given the fact that the original

costs of non-tradable shares were signi�cantly lower than the current market prices of trad-

able shares. Our measure of consideration takes into account a variety of share transfer forms

(in cash, share transfer or issuing warrant). Tradable shareholders are supposed to be better

o� with higher considerations.

The stock prices incorporate and react to a variety of information on �rm future perfor-

mance. As shown in Figure 1, positive abnormal stock returns are observed around reform

plan announcement dates, suggesting that the market expects split-share structure reform

11

plans to increase long-term �rm and equity values. We measure 3-day, 5-day and 11-day

cumulative abnormal returns, denoted by CAR3, CAR5 and CAR11 respectively, around

reform plan approval dates. We expect positive relationships between the abnormal returns

and shareholder activism.

[insert Figure 1 here]

We use the Ordinal Logit regression to examine the interaction between positive promises

and shareholder activism:

ln

�Pr (promisei)

1� Pr (promisei)

�= �+ �1 � PINi +

nXj=1

�i;j � controli;j + "i (3)

We include four types of positive promises made by �rm management, who represent non-

tradable shareholders or are non-tradable shareholders themselves. Those promises include

promise to increase shareholding by controlling shareholders (PI), promise for future dividend

payments (PD), promise for capital injection by controlling shareholders (PC) and other

types of positive promises (PO). For each promise category, we assign 1 if the �rm makes

at least one promise of the type and 0 for no promise made.3 In addition, We create a

total promise (PT ) variable that equals the sum of the four positive promise dummy values

to measure the overall quantity of positive promises made by a �rm. A larger number for

this total promise variable implies a greater bene�t for tradable shareholders in reform plan

negotiations. We expect a positive correlation between shareholder activism and positive

promises made.

[insert Table 1 here]

3Limited by our data, we are not able to measure the quality of those promises made by �rms in the stock

split share structure reform.

12

Table 1 describes the statistics of the variables in the regressions. The CAR3, CAR5 and

CAR11 are 10.3%, 10.7% and 14% respectively. Amongst control variables, we include �rm

size (SIZE), which is the natural log of �rms' assets one year prior to the split-structure

share reform. Growth (GWH) is computed using the average of the operating income growth

rates observed in the two years prior to the reform. The average growth rate is 23.7% with

a standard deviation of 56%. The oscillation is due to the fact that our sample �rms come

from 13 di�erent industries classi�ed by the CSRC. We include pro�tability (PFY ), which

is the return on �rm net assets one year before the reform. Other control variables for �rm

characteristics include market/book ratio (MB) and leverage ratio (LEV ). Yao et al: (2007)

and Zheng et al: (2007) report a positive relationship between consideration and non-tradable

share/total share ratio. Thus, we include the ratios of non-tradable shares over total shares

outstanding (RNT ) to control for that e�ect. We include �rm beta (BETA) to control for

market risk.

We use a set of dummy control variables as well. They include the batches of reform

(BATCH). The CSRC selected 30 �rms as a pilot batch to carry out split-share reforms.

Those �rms were of relatively good performance and low risk. We assign 0 to �rms that were

in the pilot batch and 1 to those were not. Firms that issue B-/H-shares may be subject to

additional shocks and in uence compared to those issue A-shares only. We therefore include

a dummy variable to specify whether a �rm issues B-/H-shares besides A-shares. A dummy

value 1 is assigned to �rms that issue B-/H-shares besides A-shares.

A Chinese mainland �rm may list its domestic A-shares on either Shanghai Stock Ex-

change or Shenzhen Stock Exchange. To control for the exchange e�ect, we include an

exchange dummy (EXCH), for which we assign 1 (0) to �rms listed on Shanghai (Shenzhen)

Stock Exchange. We include a state control dummy (SOE) as well because approximately

13

72% �rms in our sample are state-controlled. For the state control dummy, we assign 1 to

�rms that are state controlled and 0 to those are not. We include a industry dummy (IND)

to control for the industry e�ect.

Table 2 reports the correlations between the dependent variables. We �nd that considera-

tion is negatively correlated with positive promises at 1% signi�cance level. That suggests that

share transfer and positive promises could be used as substitutes to compensate non-tradable

share-holders in reform negotiations. The positive and signi�cant correlations between pairs

of positive promises indicate that �rms that made promises for dividend payments tend to

make promises for share injections as well.

[insert Table 2 here]

The cumulative abnormal returns of di�erent time window lengths tend to correlated posi-

tively and signi�cantly. We notice that cumulative abnormal returns are negatively correlated

with consideration and the total promises dummy. However, those negative relationships are

statistically insigni�cant. Since cumulative abnormal returns may be a�ected by both con-

sideration and promises, the impacts of consideration and positive promises could o�set each

other, given the negative correlation between the two. In addition, the CARs and positive

promises could be simultaneously in uenced by other factors, such as institutional owner-

ship. Therefore, the correlations reported in this table may not necessarily contradict our

cross-sectional regression results.

14

3 Regression Analysis

3.1 Main Results

Table 3 reports the regression results of the relationship between consideration and institu-

tional shareholding. As reported in Column 6, the consideration is negatively (positively)

correlated with PIN (PINSQ) with coe�cients of �2:16 (1:99). The t-statistic indicates

that those relationships are signi�cant at 1% con�dence level. Li et al: (2007) report a nega-

tive relationship between consideration and institutional shareholding as well. The evidence

indicates that there exists a nonlinear relationship between compensation value and insti-

tutional ownership. Consistent with the prediction of Hypothesis 1, the results show that

when institutional shareholders have a larger stake in a �rm, they are more likely to get

actively involved in reform negotiations for higher compensation value. Activism imposes

positive impacts on managerial decisions for non-tradable shareholders' interests. However,

evidence suggests that, when institutional ownership is relatively low, institutional investors

are prone to engaging in cooperation with �rm management in exchange for private bene�ts

that out-weight their potential gains in higher compensation value. In this case, the com-

pensation values are negatively related to institutional ownership. The results remain robust

after including control variables in our regressions.

[insert Table 3 here]

We omit share structure control variables and �rm �nancial variables in our model 1 and

2 regressions respectively. Model 3 regression includes all control variables. Amongst control

variables, the coe�cient of growth rate is�0:09 with a t-statistic of�3:95, suggesting that, for

15

�rms with greater growth opportunities, tradable shareholders are likely to demand less non-

tradable share transferred, because they could be alternatively compensated through gains

in future �rm growth. In the same vine, consideration appears negatively related with �rm

pro�tability. However, that is not statistically signi�cant. We �nd consideration is positively

correlated with state control dummy, suggesting that the non-tradable shareholders of state-

owned �rms are more willing to transfer more shares to tradable shareholders in exchange for

the trading right for their shares. One potential explanation could be that the non-tradable

shares of state-controlled �rms have lower original costs compared to those of non-state-

controlled �rms.

The non-tradable shareholders of �rms of relatively higher non-tradable share/total share

ratios transferred more shares to their tradable counterparts for compensation. The coe�-

cient of RNT is 3.28 and signi�cant at 1% level. This is consistent with the �ndings in Le

and Yuan (2006). In order to oat a large number of non-tradable shares in the market,

non-tradable shareholder are more willing to pay higher compensation to tradable sharehold-

ers. The coe�cient of 0:48 of B-/H-share dummy implies that �rms that issue B-/H-shares

pay relatively higher compensation to their tradable shareholders. Those �rms have more

incentives to send out positive signals to investors in B-/H-share markets.

Table 4 depicts the testing results for Hypothesis 2 { the relationship between positive

promises made by �rm management and institutional shareholding. Column 2 to 11 report

the regression results for total promise (PT ), promise to increase shareholding by controlling

shareholders (PI), promise for future dividend payments (PD), promise for capital injection

by controlling shareholders (PC) and other types of positive promises (PO) respectively.

[insert Table 4 here]

16

We �nd that all positive promise measures are positively correlated to institutional own-

ership. Among them, PT , PC and PO are statistically signi�cant at 1% level, and PD

is signi�cant at 5% level. The results con�rm our Hypothesis 2 that there exists overall a

positive relationship between the probability of �rm management making positive promises

in the reform and shareholder activism. Shareholder activism appears e�ective in a�ecting

managerial decisions that lead to higher �rm values in negotiations. Large non-tradable share-

holders' promises for share purchases could reduce the interest con icts between the original

non-tradable and tradable shareholders. Like Xu et al: (2008), we �nd that management's

promises for future dividend payments help to mitigate the managerial agency problems. All

evidence supports Hypothesis 2.

Table 5 illustrates the impacts of shareholder activism on the cumulative abnormal returns

around reform announcement date.

[insert Table 5 here]

We �nd that CAR3, CAR5 and CAR11 are all positively correlated with institutional

ownership. That supports our Hypothesis 3 in that the market responded more positively

to split-share structure reform plan approvals for �rms whose shares were largely held by

institutional investors. Given that market prices incorporate a variety of information about

�rm future performance and values, the results imply that institutional shareholder activism

casts positive in uence on the improvements in �rm performance.

17

3.2 Robustness Checks

To check the robustness of our �ndings, we perform the same regressions on two subsamples

{ the percentages of tradable shares held by mutual funds and quali�ed foreign institutional

investors (QFII). For mutual funds, a major institutional investor group in the Chinese

security markets, we �nd qualitatively same results as those in the main results.

Table 6 reports the regression results of the relationship between consideration and mu-

tual fund shareholding. The consideration is negatively (positively) correlated with PIN

(PINSQ) with coe�cients of �2:68 (3:34). The t-statistic indicates that the relationships

are statistically signi�cant at 1% level. The results con�rm the existence of a nonlinear rela-

tionship between compensation value and mutual fund ownership, supporting our Hypothesis

1. Furthermore, the coe�cients reported in Table 6 are of higher values compared to those

reported for the full-sample regression in Table 3, indicating that mutual fund activism was

more e�ective in pursuit for higher consideration than other institutional investor activism.

[insert Table 6 here]

Column 2 to 11 of Table 7 report the regression results for the relationships between

mutual fund activism and positive promises. We �nd that all positive promise measures

except for PI are positively correlated to mutual fund ownership. Among them, PT and PO

are statistically signi�cant at 1% level. PD and PC are signi�cant at 10% level. The results

are consistent with the ones in the full-sample regressions in that there exists an overall

positive relationship between mutual fund activism and the probability of �rm management

making positive promises.

[insert Table 7 here]

18

Table 8 reports the impacts of mutual fund activism on cumulative abnormal returns

around reform announcement dates. We �nd that CAR3, CAR5 and CAR11 are positively

correlated with mutual fund ownership. Both the coe�cients and statistical signi�cance levels

are higher compared to those in the full sample results, suggesting that the market reacted

more positively to the news of reform approvals for �rms whose shares are largely held by

mutual funds. The overall evidence indicates that mutual funds appear to play an active role

in the split-share structure reforms.

[insert Table 8 here]

In an unreported regression analysis, we perform the same tests with quali�ed foreign

institutional investor (QFII) ownerships. For all three sets of e�ectiveness measures, we

do not �nd signi�cantly positive relationships. The results suggest that quali�ed foreign

institutional investors are not in uential on Chinese companies' corporate governance.

4 Conclusions

Shareholder activism is broadly de�ned as shareholders actions to in uence managerial deci-

sions and �rm performance. In this paper, we examine the value of shareholder activism in

the split-share structure reform recently took place in China. This unique event, in nature,

allows us to avoid the de�ciencies that plague the previous literature and to present a clear

analysis on the value of shareholder activism.

The value of shareholder activism appears to be in uenced by segmented institutional

ownerships of tradable shares. Compensation values tend to be higher when institutional

shareholders hold a large percentage of tradable shares, suggesting that shareholder activism

19

plays a positive and e�ective role in a�ecting �rm management. However, compensation

values are relatively lower when institutional shareholders own a small percentage of tradable

shares. Our analysis indicates that shareholder activism could have both positive and negative

impacts in the presence of institutional investors' interest con icts.

Evidence shows that the probability of �rm management making positive promises is

positively correlated to institutional shareholder activism, o�ering support to the argument

of e�ective shareholder activism. Stock prices react positively to institutional shareholder

activism around reform approvals as well. We report both economically and statistically sig-

ni�cant positive relationships between abnormal stock returns and institutional shareholding.

Shareholder activism in China is still in its primitive stage. Nevertheless, the split-share

structure reform provides Chinese institutional shareholders an opportunity to get actively

involved in �rm management. In this special event, they have demonstrated their strength in

a�ecting managerial decisions and �rm future performance. They are expected to play more

signi�cant roles in corporate governance.

20

ReferenceAllen, F., Qian, J. & Qian, M. (2005), "Law, �nance, and economic growth in China",

Journal of Financial Economics, 77: 57-116.

Carvell, S. A. & Strebel, P. J. (1987), "Is There A Neglected Firm E�ect?", Journal of

Business Finance & Accounting 14: 279-290.

Carleton, W. T., Nelson, J. M., & Weisbach, M. S. (1998), "The in uence of institutions

on corporate governance through private negotiations: Evidence from TIAA-CREF", Journal

of Finance 53: 1335-1362.

Del Guercio, D. & Hawkins, J. (1999), "The Motivation and Impact of Pension Fund

Activism", Journal of Financial Economics 52: 293-340.

Edelman, R. B. & Baker H. K. (1990), "Liquidity and Stock Exchange Listing", The

Financial Review 25: 231-249.

Feng, L. & Xu, W. (2006), "The Abnormal Returns of Pilot Firms in the Split-share

Structure Reform", Contemporary Fiance & Economics 2.

Gillan, S. L. & Starks, L. T. (1998), "A Survey of Shareholder Activism: Motivation and

Empirical Evidence", Contemporary Finance Digest 2: 10-34.

Gillan, S. L. & Starks, L. T. (2000), "Relationship investing and shareholder activism by

institutional investors", Journal of Financial Economics 57: 275

Hirschman, A. (1971), "Exit, Voice and Loyalty: Responses to Decline in Firms, Organi-

zations, and States", Harvard University Press.

Hwang, C., Zhang S. & Zhu, Y. (2006), "Float Liquidity, Speculation, and Stock Prices:

Evidence from the Share Structure Reform in China",Working Paper, Nanyang Technological

University.

21

Karpo�, J. M. (2001), "The Impact of Shareholder Activism on Target Companies: A

Survey of Empirical Findings", Working Paper, University of Washington.

Karpo�, J. M., Malatesta, P. H. & Walking, R. A. (1996), "Corporate Governance and

Shareholder Initiatives: Empirical Evidence", Journal of Financial Economics 42: 365-395.

Le, Q. & Yuan, H. (2006), "Do Compensation Raitos Re ect Firm Pro�tability and

Risk?", China Accounting & Financial Studies 8.

Li, K., Wang, T., Cheung, Y. & Jiang, P. (2007), "Privatization with Negotiation: Evi-

dence from the Split Share Structure Reform in China",Working Paper, University of British

Columbia.

Liao, L., Li, J., Liu, B. & Wang, H. (2008), "Price Discovery and Market E�ciency: Evi-

dence from the Split-Share Structure Reform in China",Working Paper, Tsinghua University.

McConnel, J. J. & Servaes, H. (1995), "Equity Ownership and the Two Faces of Debt",

Journal of Financial Economics 39: 131-157.

Pound, J. (1988), "Proxy Contest and the E�ciency of Shareholder Oversight", Journal

of Financial Economics 20: 237-265.

Smith, M. P. (1996), "Shareholder Activism by Institutional Investors: Evidence from

CALPERS", Journal of Finance 51: 227-252.

Strickland, D., Wiles, K. & Zenner, M. (1996), "A requiem for the USA: Is small share-

holder monitoring e�ective?" Journal of Financial Economics 40: 319-338.

Wahal, S. (1996), "Pension Fund Activism and Firm Performance", Journal of Financial

and Quantitative Analysis 31: 1-23.

Woidtke, T. (2002), "Agents Watching Agents?: Evidence from Pension Fund Ownership

and Firm Value", Journal of Financial Economics 63: 99-131.

Xu, L. (2006), "A Study on the Transactions of the Multiple Accounts of Institutional

22

Investors", Economic Studies 2.

Xu, N., Zhang, H. & Wu, S. (2008), "Are there signaling e�ects of �rm promises?",

Management World 3.

Yao, Y., Liu, Z. & Wang, J. (2007), "Split-share Structure Reform, Institutional Investors

and Investor Protection", Financial Studies 11.

Zheng, Z., Sun, Y., Tan, S. & Jiang, D. (2007), "The Impacts of Split-share Structure

Reform on Corporate Governance of Publicly Listed Companies in China", Economic Studies

7.

23

Table 1: Descriptive Statistics of Regression Variables

This table reports the descriptive statistics of regression variables. CONS denotes consideration. PT denotes total positive promises.PI denotes promise for equity purchase. PD denotes promise for dividend payments. PC denotes promise for capital injection.CAR3; 5; 11 denote 3-,5-11-day cumulative abnormal returns respectively. PIN denotes the percentage of tradable shares held byinstitutional investors. PFY denotes pro�tability. BETA denotes �rm CAPM �. LEV denotes leverage ratio. GWH denotes growth.SOE denotes state-controlled. RNT denotes non-tradable/total share ratio. BH denotes issuing B-/H-shares. SIZE denotes log of�rm size. EXCH denotes exchange. BATCH denotes reform batch.

Mnemonic Mean Standard Deviation Median Min P5 P95 Max

CONS 2.960 0.789 3.000 0.220 1.379 4.000 7.000PT 0.465 0.674 0.000 0.000 0.000 2.000 3.000PI 0.076 0.265 0.000 0.000 0.000 1.000 1.000PD 0.222 0.416 0.000 0.000 0.000 1.000 1.000PC 0.044 0.205 0.000 0.000 0.000 0.000 1.000PO 0.123 0.329 0.000 0.000 0.000 1.000 1.000CAR3 0.103 0.259 0.052 -0.359 -0.117 0.468 3.231CAR5 0.107 0.265 0.060 -0.359 -0.134 0.479 3.249CAR11 0.140 0.280 0.085 -0.344 -0.135 0.579 3.233PIN 0.094 0.142 0.024 0.000 0.000 0.441 0.861PFY 0.004 0.267 0.041 -4.830 -0.314 0.194 0.410BETA 1.074 0.302 1.062 0.071 0.592 1.586 2.233LEV 0.503 0.182 0.512 0.013 0.184 0.787 1.173GWH 0.237 0.560 0.188 -0.873 -0.211 0.768 14.559SOE 0.720 0.449 1.000 0.000 0.000 1.000 1.000RNT 0.605 0.112 0.622 0.146 0.392 0.750 0.913BH 0.093 0.291 0.000 0.000 0.000 1.000 1.000SIZE 21.326 0.967 21.260 18.324 19.929 22.951 26.978EXCH 0.646 0.479 1.000 0.000 0.000 1.000 1.000BATCH 0.973 0.163 1.000 0.000 1.000 1.000 1.000year2006 0.784 0.411 1.000 0.000 0.000 1.000 1.000year2007 0.065 0.246 0.000 0.000 0.000 1.000 1.000

24

Table 2: Correlations of Dependent Variables

This table reports the correlations of dependent variables. CONS denotes consideration. PT denotes total positive promises. PIdenotes promise for equity purchase. PD denotes promise for dividend payments. PC denotes promise for capital injection. CAR3; 5; 11denote 3-,5-11-day cumulative abnormal returns respectively. Note: *, ** and *** represent 10%, 5% and 1% signi�cance levels respectively(two-tails).

CONS PT PI PD PC PO CAR3 CAR5 CARS11CONS 1PT -0.337*** 1PI -0.154*** 0.509*** 1PD -0.125*** 0.682*** 0.104*** 1PC -0.132*** 0.349*** 0.023 0.025 1PO -0.325*** 0.558*** 0.092*** 0.034 0.042 1CAR3 -0.059* -0.015 -0.051* -0.037 0.02 0.046 1CAR5 -0.052* -0.014 -0.053* -0.036 0.024 0.045 0.985*** 1CAR11 -0.037 -0.019 -0.057* -0.032 0.023 0.034 0.924*** 0.945*** 1

25

Table 3: Regression Results for consideration

This table reports the regression results for consideration. PIN denotes the percentage of tradable shares held by institutionalinvestors. PINSQ denotes the square of the percentage of tradable shares held by institutional investors. PFY denotes pro�tability.BETA denotes �rm CAPM �. LEV denotes leverage ratio. GWH denotes growth. SOE denotes state-controlled. RNT denotesnon-tradable/total share ratio. BH denotes issuing B-/H-shares. SIZE denotes log of �rm size. EXCH denotes exchange. BATCHdenotes reform batch. Note: *, ** and *** represent 10%, 5% and 1% signi�cance levels respectively (two-tails).

Model 1 Model 2 Model 3Coe¢ cient t-value Coe¢ cient t-value Coe¢ cient t-value

Intercept 2.988*** 4.50 1.847*** 3.06 1.636*** 2.62PIN -1.947*** -4.41 -2.417*** -6.05 -2.162*** -5.44PINSQ 1.518* 1.89 2.220*** 2.89 1.991*** 2.63PFY 0.018 0.080 -0.11 -0.53BETA 0.247*** 2.66 0.151* 1.81LEV 0.051 0.35 0.11 0.91GWH -0.139*** -4.85 -0.0915*** -3.95SOE 0.374*** 7.36 0.379*** 7.68RNT 3.310*** 15.95 3.285*** 15.82BH 0.520*** 5.75 0.482*** 5.99SIZE -0.004 -0.15 -0.04 -1.40 -0.04 -1.27EXCH 0.076 1.62 -0.01 -0.23 -0.01 -0.23BATCH -0.16 -0.92 -0.11 -0.70 -0.11 -0.68Y ear2006 -0.351*** -5.50 -0.306*** -5.46 -0.332*** -5.76Y ear2007 -0.696*** -4.96 -0.581*** -5.27 -0.677*** -5.23Industry As control variable As control variable As control variable

N 1095 1095 1095Adj. R-sq 0.115 0.346 0.354

26

Table 4: Regression Results for Positive Promises

This table reports the regression results for positive promises. PIN denotes the percentage of tradable shares held by institutionalinvestors. PFY denotes pro�tability. BETA denotes �rm CAPM �. LEV denotes leverage ratio. GWH denotes growth. SOE denotesstate-controlled. RNT denotes non-tradable/total share ratio. BH denotes issuing B-/H-shares. SIZE denotes log of �rm size. EXCHdenotes exchange. BATCH denotes reform batch. PI denotes promise to increase shareholdings by controlling shareholders. PD denotespromise for future dividend payments. PC denotes promise for capital injection by controlling shareholders. PO denotes other types ofpositive promises, and PT denotes total promise. Note: *, ** and *** represent 10%, 5% and 1% signi�cance levels respectively (two-tails).

M o d e l 1 M o d e l 2 M o d e l 3 M o d e l 4 M o d e l 5

PT PI PD PC POC o e ¢ c i e n t t - va lu e C o e ¢ c i e n t t - va lu e C o e ¢ c i e n t t - va lu e C o e ¢ c i e n t t - va lu e C o e ¢ c i e n t t - va lu e

Intercept - 6 .7 9 6 * - 1 .6 8 - 1 1 .4 3 * * * - 4 .7 - 1 0 .3 1 * * * - 2 .7 5 1 .8 9 4 0 .6 1

PIN 2 .4 7 5 * * * 5 .0 9 0 .1 1 4 0 .1 3 1 .3 2 4 * * 2 .2 8 2 .0 8 6 * * * 2 .8 5 2 .9 4 7 * * * 4 .3 3

PFY - 0 .4 3 3 - 1 .2 4 0 .7 6 3 0 .5 9 2 .5 2 0 * * 2 .3 9 - 0 .8 6 5 * * - 2 .5 - 0 .2 3 6 - 0 .7 5

BETA - 0 .2 6 - 1 .1 6 - 0 .4 6 5 - 1 .0 1 - 0 .3 1 2 - 1 .0 9 0 .5 5 3 0 .9 - 0 .1 0 1 - 0 .2 9

LEV - 1 .6 7 0 * * * - 3 .9 4 - 1 .4 5 1 * - 1 .6 7 - 2 .6 5 8 * * * - 5 .2 2 - 1 .5 0 4 - 1 .5 8 1 .0 1 3 1 .4 7

GWH 0 .1 9 6 * * * 3 .3 9 0 .3 1 7 * * * 2 .7 1 0 .3 3 9 * * 2 .2 1 - 0 .0 2 2 - 0 .1 7 - 0 .4 2 - 1 .1 5

SOE - 0 .6 2 3 * * * - 3 .9 - 1 .0 8 3 * * * - 3 .7 2 - 0 .1 2 3 - 0 .6 4 0 .6 0 1 1 .3 6 - 1 .2 0 3 * * * - 5 .5

RNT - 2 .2 5 6 * * * - 3 .5 3 - 4 .9 7 6 * * * - 4 .5 8 - 0 .8 9 6 - 1 .1 8 - 0 .9 7 1 - 0 .7 6 - 3 .3 2 9 * * * - 3 .8 1

BH - 0 .6 7 6 * * * - 2 .6 3 - 0 .2 1 1 - 0 .4 3 - 0 .8 7 1 * * * - 2 .6 5 0 .0 2 6 2 0 .0 6 - 0 .8 9 4 * * - 2 .1 7

SIZE 0 .3 7 2 * * * 4 .2 6 0 .5 2 9 * * * 2 .8 1 0 .4 6 7 * * * 4 .3 3 0 .2 4 8 1 .3 7 - 0 .0 9 0 2 - 0 .6 6

EXCH - 0 .0 3 7 4 - 0 .2 8 0 .1 0 5 0 .3 9 0 .2 3 8 1 .4 4 0 .2 8 4 0 .8 3 - 0 .4 2 3 * * - 2 .1

BATCH - 0 .5 1 3 - 1 .4 9 - 2 .3 4 3 * * * - 4 .4 1 .4 3 6 * * * 2 .7 2 0 .1 5 9 0 .2 2

Y ear2006 0 .0 7 2 0 .3 7 - 0 .7 3 6 * * - 2 .1 9 0 .3 0 9 1 .3 2 1 .1 4 5 1 .5 7 0 .2 1 1 0 .6 5

Y ear2007 0 .1 3 2 0 .4 0 .0 9 2 1 0 .2 2 .1 9 5 * * * 2 .6 5 0 .1 8 7 0 .3 6

Industry A s c o n t r o l va r ia b le A s c o n t r o l va r ia b l e A s c o n t r o l va r ia b l e A s c o n t r o l va r ia b l e A s c o n t r o l va r ia b l e

N 1 0 9 5 1 0 0 2 1 0 9 5 1 0 1 4 1 0 5 0

p s e u d o R - s q 0 .0 6 9 0 .2 1 2 0 .1 2 2 0 .1 0 5 0 .1 2 5

27

Table 5: Regression Results for Cumulative Abnormal Returns around Reform Plan Approval Day

This table reports the regression results for cumulative abnormal returns. PIN denotes the percentage of tradable shares held byinstitutional investors. PFY denotes pro�tability. BETA denotes �rm CAPM �. LEV denotes leverage ratio. GWH denotes growth.SOE denotes state-controlled. RNT denotes non-tradable/total share ratio. BH denotes issuing B-/H-shares. SIZE denotes log of�rm size. EXCH denotes exchange. BATCH denotes reform batch. Note: *, ** and *** represent 10%, 5% and 1% signi�cance levelsrespectively (two-tails).

Model 1 Model 2 Model 3CAR3 CAR5 CAR11

Coe¢ cient t-value Coe¢ cient t-value Coe¢ cient t-valueIntercept 0.318 1.53 0.311 1.5 0.348 1.64PIN 0.0862* 1.93 0.0847* 1.89 0.0899* 1.94SIZE -0.0153 -1.52 -0.0151 -1.5 -0.0169 -1.64MB -0.00438 -1.37 -0.00493* -1.69 -0.00440* -1.83LEV 0.0629 1.29 0.0754 1.53 0.0841* 1.67BH 0.0947** 2.4 0.0949** 2.5 0.0897** 2.41

Y ear2006 0.0475*** 4.34 0.0488*** 4.37 0.0790*** 6.53Y ear2007 0.554*** 7.68 0.574*** 8.01 0.633*** 8.89

N 1095 1095 1095Adj. R-sq 0.265 0.274 0.282

28

Table 6: Regression Results for consideration (Mutual Funds)

This table reports the regression results for consideration (mutual funds). PIN denotes the percentage of tradable shares held byinstitutional investors. PINSQ denotes the square of the percentage of tradable shares held by institutional investors. PFY denotespro�tability. BETA denotes �rm CAPM �. LEV denotes leverage ratio. GWH denotes growth. SOE denotes state-controlled. RNTdenotes non-tradable/total share ratio. BH denotes issuing B-/H-shares. SIZE denotes log of �rm size. EXCH denotes exchange.BATCH denotes reform batch. Note: *, ** and *** represent 10%, 5% and 1% signi�cance levels respectively (two-tails).

Model 1 Model 2 Model 3Coe¢ cient t-value Coe¢ cient t-value Coe¢ cient t-value

Intercept 3.152*** 4.8 1.972*** 3.33 1.748*** 2.86PIN -2.222*** -4.38 -3.024*** -7.07 -2.675*** -6.23PINSQ 2.469** 2.34 3.852*** 4.14 3.446*** 3.79PFY 0.000372 0 -0.127 -0.6BETA 0.263*** 2.8 0.162* 1.93LEV 0.0847 0.58 0.145 1.15GWH -0.139*** -4.78 -0.0898*** -3.86SOE 0.384*** 7.51 0.389*** 7.85RNT 3.328*** 15.92 3.302*** 15.84BH 0.525*** 5.7 0.484*** 5.94SIZE -0.0155 -0.53 -0.0456* -1.77 -0.0447 -1.64EXCH 0.0686 1.45 -0.017 -0.43 -0.0161 -0.41BATCH -0.172 -0.95 -0.119 -0.76 -0.116 -0.73Y ear2006 -0.343*** -5.45 -0.296*** -5.31 -0.325*** -5.68Y ear2007 -0.694*** -4.92 -0.566*** -5.08 -0.675*** -5.17Industry As control variable As control variable As control variable

N 1095 1095 1095adj. R-sq 0.104 0.338 0.347

29

Table 7: Regression Results for Positive Promises (Mutual Funds)

This table reports the regression results for positive promises (mutual funds). PIN denotes the percentage of tradable shares heldby institutional investors. PFY denotes pro�tability. BETA denotes �rm CAPM �. LEV denotes leverage ratio. GWH denotesgrowth. SOE denotes state-controlled. RNT denotes non-tradable/total share ratio. BH denotes issuing B-/H-shares. SIZE denoteslog of �rm size. EXCH denotes exchange. BATCH denotes reform batch. PI denotes promise to increase shareholdings by controllingshareholders. PD denotes promise for future dividend payments. PC denotes promise for capital injection by controlling shareholders.PO denotes other types of positive promises, and PT denotes total promise. Note: *, ** and *** represent 10%, 5% and 1% signi�cancelevels respectively (two-tails).

M o d e l 1 M o d e l 2 M o d e l 3 M o d e l 4 M o d e l 5

PT PI PD PC POC o e ¢ c i e n t t - va lu e C o e ¢ c i e n t t - va lu e C o e ¢ c i e n t t - va lu e C o e ¢ c i e n t t - va lu e C o e ¢ c i e n t t - va lu e

Intercept - 7 .1 1 8 * - 1 .7 8 - 1 1 .6 7 * * * - 4 .7 9 - 1 0 .6 9 * * * - 2 .8 4 1 .7 8 9 0 .5 7

PIN 2 .5 2 1 * * * 4 .4 2 - 0 .5 5 3 - 0 .5 3 1 .1 5 9 * 1 .6 6 1 .9 0 5 * 1 .9 4 3 .6 2 1 * * * 4 .4 3

PFY - 0 .3 8 - 1 .0 4 1 .0 3 8 0 .7 5 2 .7 3 6 * * 2 .5 2 - 0 .8 3 3 * * - 2 .4 - 0 .2 0 6 - 0 .6 5

BETA - 0 .2 9 7 - 1 .3 1 - 0 .5 0 6 - 1 .1 - 0 .3 3 4 - 1 .1 7 0 .5 1 8 0 .8 4 - 0 .1 1 - 0 .3 1

LEV - 1 .7 0 5 * * * - 4 .0 4 - 1 .5 1 3 * - 1 .7 4 - 2 .6 8 4 * * * - 5 .2 8 - 1 .5 7 0 * - 1 .6 9 1 .0 3 1 1 .5 1

GWH 0 .1 9 7 * * * 3 .4 4 0 .3 1 8 * * * 2 .7 2 0 .3 4 2 * * 2 .1 8 - 0 .0 1 2 3 - 0 .1 - 0 .4 3 3 - 1 .1 7

SOE - 0 .6 3 8 * * * - 3 .9 9 - 1 .0 8 1 * * * - 3 .7 - 0 .1 2 8 - 0 .6 7 0 .5 8 8 1 .3 5 - 1 .2 2 8 * * * - 5 .6 1

RNT - 2 .2 7 5 * * * - 3 .5 6 - 4 .9 6 3 * * * - 4 .5 6 - 0 .9 1 6 - 1 .2 1 - 1 .0 1 9 - 0 .8 1 - 3 .4 0 0 * * * - 3 .8 6

BH - 0 .6 4 7 * * - 2 .5 5 - 0 .2 1 2 - 0 .4 4 - 0 .8 6 4 * * * - 2 .6 3 0 .0 3 0 5 0 .0 7 - 0 .8 7 4 * * - 2 .1 3

SIZE 0 .3 9 5 * * * 4 .5 1 0 .5 4 8 * * * 2 .9 5 0 .4 8 2 * * * 4 .4 7 0 .2 7 3 1 .5 2 - 0 .0 8 5 4 - 0 .6 2

EXCH - 0 .0 2 4 3 - 0 .1 8 0 .1 1 6 0 .4 3 0 .2 4 7 1 .5 0 .3 0 4 0 .8 9 - 0 .4 0 5 * * - 2 .0 1

BATCH - 0 .4 8 - 1 .4 1 - 2 .3 6 8 * * * - 4 .4 6 1 .4 3 5 * * * 2 .7 4 0 .2 5 4 0 .3 3

Y ear2006 0 .0 6 1 9 0 .3 2 - 0 .7 2 4 * * - 2 .1 4 0 .3 1 4 1 .3 4 1 .1 5 1 .5 8 0 .2 0 6 0 .6 3

Y ear2007 0 .1 2 7 0 .3 9 0 .1 0 1 0 .2 2 2 .1 9 1 * * * 2 .6 4 0 .1 8 5 0 .3 6

Industry A s c o n t r o l va r ia b le A s c o n t r o l va r ia b l e A s c o n t r o l va r ia b l e A s c o n t r o l va r ia b l e A s c o n t r o l va r ia b l e

N 1 0 9 5 1 0 0 2 1 0 9 5 1 0 1 4 1 0 5 0

p s e u d o R - s q 0 .0 6 5 0 .2 1 2 0 .1 2 0 .1 0 1 0 .1 2 6

30

Table 8: Regression Results for Cumulative Abnormal Returns around Reform Plan Approval Day(Mutual Funds)

This table reports the regression results for cumulative abnormal returns (mutual funds). PIN denotes the percentage of tradableshares held by institutional investors. PFY denotes pro�tability. BETA denotes �rm CAPM �. LEV denotes leverage ratio. GWHdenotes growth. SOE denotes state-controlled. RNT denotes non-tradable/total share ratio. BH denotes issuing B-/H-shares. SIZEdenotes log of �rm size. EXCH denotes exchange. BATCH denotes reform batch. Note: *, ** and *** represent 10%, 5% and 1%signi�cance levels respectively (two-tails).

Model 1 Model 2 Model 3CAR3 CAR5 CAR11

Coe¢ cient t-value Coe¢ cient t-value Coe¢ cient t-valueIntercept 0.326 1.58 0.326 1.58 0.355* 1.68PIN 0.116** 2.19 0.122** 2.3 0.118** 2.14SIZE -0.0157 -1.57 -0.0158 -1.58 -0.0171* -1.67MB -0.00458 -1.43 -0.00521* -1.78 -0.00458* -1.89LEV 0.0642 1.32 0.0777 1.59 0.0851* 1.7BH 0.0953** 2.41 0.0958** 2.52 0.0903** 2.43

Y ear2006 0.0475*** 4.34 0.0489*** 4.37 0.0791*** 6.52Y ear2007 0.554*** 7.68 0.574*** 8.01 0.633*** 8.89

N 1095 1095 1095Adj. R-sq 0.265 0.275 0.282

31

Figure 1: Abnormal Returns and Cumulative Abnormal Returns around Refrom Approval

The �gure illustrates the averages abnormal returns (ARs) and cumulative abnormal returns (CARs) around reform plan approvalsfor �rms in our sample. The CARs reported in Figure 1(B) are computed using the abnormal return on day -10 as base. The CARs usedin our regressions are computed using the abnormal returns around the reform plan approval (announcement) dates. For example CAR3equals the sum of AR�1, AR0 and AR1.

02

46

810

12Ab

norm

al R

etur

n %

­10 ­5 0 5 10Final Announcement Day

Abnormal Returns around Final Announcement Day

(A) Abnormal Returns

03

69

1215

1821

Cum

ulat

ive 

Abno

rmal

 Ret

urn 

%

­10 ­5 0 5 10Final Announcement Day

CARs around Final Announcement Day

(B) Cumulative Abnormal Returns

32